A comprehensive hybrid modeling approach was developed, combining Aspen Plus for system-level process simulations with high-fidelity CFD models that resolve reacting flow and combustion behavior. Detailed chemical kinetic mechanisms were employed to capture pollutant formation pathways across a wide range of multi-fuel compositions.
Outputs from the simulations were used to train machine learning algorithms capable of rapidly exploring the design space and identifying fuel mixtures that offer the best trade-offs between emissions, cost, and performance. The methodology is applicable to both air-fired and oxy-combustion environments, and supports the integration of carbon capture technologies.
This work demonstrates how advanced modeling and AI can accelerate cleaner combustion strategies for legacy energy systems. The results provide insight into pollutant chemistry from complex fuel blends and enable rational fuel design for modernized, lower-emission high-temperature processes.